Certified 100% AI-free organic content • Hacker News Board

Re: Certified 100% AI-free organic content

#171

Post by alpos » Tue, Jan 24, 2023, 5:13 PM UTC

Veuxdo wrote: post #2

"AI-free" is pretty clear. "Organic" is much more subjective. The same applies to Organic food, if I'm not mistaken.

Actually it's the other way around.

Those pushing the concept of "AI-Free" have yet to nail down what amounts and types of automation may be used in the production of a thing before it can no longer be labeled "AI-Free".

On the other hand, "Organic" is very well defined at this point. In the US there is a whole legal framework around labeling foods, drugs, and cosmetics.

https://www.fda.gov/food/food-labeling-nutrition/organic-foo...

https://www.ams.usda.gov/about-ams/programs-offices/national...

There is even a training and accreditation system in place for qualifying people as certifiers of organic practices as well as an application and review process for farms and manufacturers to be certified to use the label.

https://www.ams.usda.gov/services/organic-certification/beco...

Re: Certified 100% AI-free organic content

#172

Post by felideon » Tue, Jan 24, 2023, 5:21 PM UTC

artpi wrote: post #7

dsign wrote: post #4

There is so much stuffing for a simple idea that I'm not sure if this piece deserves its own title, but I'll give it the benefit of the doubt. One thing that I wonder though is how we will draw the line. If I'm writing a piece and do a Google search, and in that way invoke BERT under the hood, is anything that I write afterwards "AI-tainted"? What about the grammar checker? Or the spot removal tool in photoshop or gi…

> There is so much stuffing for a simple idea that I'm not sure if this piece deserves its own title, but I'll give it the benefit of the doubt. Frankly I had the same thought writing it :D It's more of a stake in the ground sort of a thing I guess? What I really want is somebody saying "hey, there is an open standard already here" so I can use it.

By contrast, I enjoyed the entire piece, read another one of your posts, and subscribed to your newsletter.

Re: Certified 100% AI-free organic content

#173

Post by DennisP » Tue, Jan 24, 2023, 5:21 PM UTC

luxcem wrote: post #152

dale_glass wrote: post #15

> Published content will be later used to train subsequent models, and being able to distinguish AI from human input may be very valuable going forward I find this to be a particularly interesting problem in this whole debacle. Could we end up having AI quality trend downwards due to AI ingesting its own old outputs and reinforcing bad habits? I think it's a particular risk for text generation. I've already run into…

> Published content will be later used to train subsequent models, and being able to distinguish AI from human input may be very valuable going forward Every discussion on AI take the example of ChatGPT and its inherent flaws but AI-generated content doesn't have to be dull and low quality. One question that bother me is does it really matter? If AI-generated content is on par with Human-made or even better does it m…

I don't think AI has to be low-quality for GP's concern to be valid.

Humans get inputs from a large variety of sources, but if an AI's input is just text, then there's the potential for AI's input to mostly consist of its prior output. Iterate this, and its model could gradually diverge from the real world.

The equivalent in human society is groupthink, where members of a subgroup get most of their information from the same subgroup and end up believing weird things. We can counter that by purposely getting inputs from outside of our group. For a text-model AI, this means identifying text that wasn't produced by AI, as the article suggests.

Re: Certified 100% AI-free organic content

#174

Post by CuriouslyC » Tue, Jan 24, 2023, 5:25 PM UTC

rcme wrote: post #163

CuriouslyC wrote: post #144

Earlier quoted context omitted.

Not necessarily. The bloom model (a GPT competitor and similarly sized) was trained on 1.5T of text, which reduces down to 350B unique tokens. If you took a histogram of those unique tokens, it would have a very long tail with probably 1% or less being well represented. That leaves 350M common tokens to serve as the basis for token tuples being fed into crowdsourcing. There are probably ~2-5B very common token sequen…

BLOOM isn’t as good as GPT-3 because it doesn’t use as much training data. LLM quality is still data bound [0]. Further limiting data by requiring annotation is not going to work, at least with the current LLM modeling approach. 0: https://www.alignmentforum.org/posts/6Fpvch8RR29qLEWNH/chinc...

As the scale of input data goes up linearly, the scale of commonly observed input patterns goes up logarithmically. If we bumped the scale up an order of magnitude in terms of common input tokens, that still means we could annotate the important part of a 150TB text corpus for 125B worth of human annotation. Given that could break the budget of even large corporations, realistically we'd probably train a model to predict the scores of interest using a fraction of that much human annotation, which would be inferior but still a massive improvement. It is also likely that corporations would team up with indirect competitors to share the cost of annotation and gain an advantage against direct competitors.

Re: Certified 100% AI-free organic content

#175

Post by throwaway589275 » Tue, Jan 24, 2023, 5:27 PM UTC

icepat wrote: post #135

throwaway589275 wrote: post #58

Earlier quoted context omitted.

> KiwiFarms is just a website to organize doxxing and online harassment KF is a gossip site, like Lipstick Alley or DataLounge, but with a focus on e-celebs and internet phenomena. I go there whenever I want to get the dirt (ie. truth) about e-celebs or some weird internet trend or subculture. Its users literally just passively document and discuss this stuff, and there's a policy (and ethos) of "no touch." You and o…

The sorts of people being "documented" on KiwiFarms are not celebrities. They're usually vulnerable people with some sort of mental illness who are struggling. And I don't buy for a moment the "no touch" policy. Just because you can't use a specific website to harass someone, does not mean you can't use the information on the website to harass them off-platform. This is a bad take. There's a major difference between…

[flagged]

Re: Certified 100% AI-free organic content

#176

Post by seydor » Tue, Jan 24, 2023, 5:28 PM UTC

There is a case to be made that talking to a human is the bland one now. Amidst all the censoring and self-sensoring and assumption of bad faith, there is a lot that is not possible to discuss in polite company, even online, but is interesting. AIs seem more willing to give me a straight answer rather than going of moral tangents or ego trips. I have had discussions with the chat that went deeper than the smartasses in reddit . I thought that that was a lost art

So, just like with good cheese, maybe i will go with full-fat

Re: Certified 100% AI-free organic content

#177

Post by BashiBazouk » Tue, Jan 24, 2023, 5:28 PM UTC

Reminds me of the Dovetail phyle in Neal Stephenson's The Diamond Age where the rich value purely human made goods in a world of matter compilers that can almost instantly make or duplicate just about anything.

On the other hand, I find Stable Diffusion the most interesting thing going on in art at the moment...

Re: Certified 100% AI-free organic content

#178

Post by alpos » Tue, Jan 24, 2023, 5:28 PM UTC

layer8 wrote: post #164

alpos wrote: post #133

Earlier quoted context omitted.

This take strikes me as a little off. Programming languages are language. Unlike natural languages they are also based on context-free grammar. So an understanding of programming languages should actually be easier for even a general language model to incorporate than natural languages. We can expect a bot like this to not really get context clues in natural language, although they seem to be getting better at that,…

The failure to realize that the API doesn’t exist and therefore the code won’t work in practice, however, is a major lack of insight and understanding.

Agreed. That does seem to be an example of a language model failing to understand the context in which the question was asked.

The user was implicitly expecting code that would function when executed immediately and as written with no additional supporting libraries included. This is different from code that would function correctly when executed after having downloaded relevant existing packages. Which is different from code that would function if executed along side additional supporting code from private libraries the user might not have access to. Etc...

Yet any of those answers fit for the same prompt, "create an R script that does such and such". The bot's lack of insight is on the likely intention behind the prompt rather than on the requested language. I'd say if it produces any code that fits the syntax and grammatical structure of the requested language, that's enough to say it understands the language.

Re: Certified 100% AI-free organic content

#179

Post by rosywoozlechan » Tue, Jan 24, 2023, 5:28 PM UTC

> do you really want to cry watching a movie that was 100% produced by robots?

I've had a lot of fun having ChatGPT write stories for me: I'd ask it make changes, to add a character, add a motivation, etc. I'm just playing around, and it's 100% produced by a robot and I enjoy it. I don't personally mind having an emotional response by a story generated by a "robot". I don't really understand how it being bot generated cheapens the experience. The emotion that I feel are elicited by my thoughts and reflections based on what I've read and experienced, not by the robot.

Re: Certified 100% AI-free organic content

#180

Post by TylerE » Tue, Jan 24, 2023, 5:29 PM UTC

rgrieselhuber wrote: post #95

dazc wrote: post #86

Earlier quoted context omitted.

> 'It would be very interesting if companies and site owners wished to self-certify their site as organic with something like an /ai-free.txt.' I'm sure you can appreciate that such an initiative would be wholesale abused from day 1.

As with anything in the industry, yes. But it would provide a basis for ranking penalties if the search engines cared about it.

I'd bet money on it being an INVERSE signal.